Running Head: CHILDREN’S SPELLING IN ENGLISH AND PORTUGUESEspell.psychology.wustl.edu/PolloVSyllLN/PolloVSyllLN.pdf · Spanish, and Kamii et al. pointed to the relative unclearness

Vowels, syllables, and letter names 1

Running Head: CHILDREN’S SPELLING IN ENGLISH AND PORTUGUESE

Vowels, Syllables, and Letter Names: Differences Between Young Children’s Spelling in

English and Portuguese

Tatiana Cury Pollo

Brett Kessler

Rebecca Treiman

Washington University in St. Louis


Abstract

Young Portuguese-speaking children have been reported to produce more vowel- and

syllable-oriented spellings than English speakers. To investigate the extent and source of such

differences, we analyzed children’s vocabulary and found that Portuguese words have more

vowel letter names and a higher vowel–consonant ratio than English. In a spelling experiment,

we found that Portuguese speakers used more vowels than English speakers but did not produce

more syllabic spellings. The differences that we observed are attributable to quantitative

differences in the languages and their writing and letter name systems; they do not support the

widespread idea that speakers of Romance languages pass through an additional, syllabic, stage

of development.


Vowels, Syllables, and Letter Names: Differences Between Young Children’s Spelling in

English and Portuguese

Because most research on reading and spelling has studied English, our understanding of

literacy and its development may be constrained by the characteristics of English and its writing

system, which may not generalize to other languages. In an attempt to broaden our knowledge,

researchers have recently turned to cross-linguistic studies. This research (e.g., Aro & Wimmer,

2003; Defior, Martos, & Cary, 2002; Seymour, Aro, & Erskine, 2003) has tended to focus on

reading. Our goal in the present study is to extend the investigation to spelling by examining the

early acquisition of spelling in two writing systems, those of Brazilian Portuguese and American

English. Several differences have been reported between spellings produced by English speakers

and those produced by speakers of Romance languages such as Portuguese, Spanish, and Italian.

We address here two of the most prominent claims: that early spellers in the Romance languages

use a disproportionate amount of vowels, often omitting consonants entirely; and that they tend

to spell syllabically, writing one symbol per syllable. Such observations are rarely adduced for

early writers of English. We describe new studies designed to determine whether there are

indeed cross-linguistic differences along these dimensions, and, if so, to what extent such

differences can be explained by differences in the languages themselves or in their writing

systems.

Among English-speaking children, vowels are a major locus of children’s spelling

mistakes (Read, 1986; Treiman, 1993; Varnhagen, Boechler, & Steffler, 1999). Kamii, Long,

Manning, and Manning (1990) reported that English-speaking kindergartners sometimes

represent only consonants in their spellings. The opposite pattern is more commonly reported

among speakers of Romance languages, where all-vowel spellings are often cited, such as AO for


Spanish sapo ‘frog’ (Ferreiro & Teberosky, 1982), and UUU for Portuguese urubu ‘vulture’

(Nunes Carraher & Rego, 1984). Although little quantitative data exist, the reports suggest some

important differences between use of consonants and vowels in English and in other languages.

Another apparent difference between early spelling in Romance languages and English

involves syllabic spellings. Ferreiro and colleagues (Ferreiro, 1990; Ferreiro & Teberosky, 1982;

Vernon & Ferreiro, 1999) reported that young Spanish-speaking children often write one symbol

per syllable. Reports of syllabic spellings among preschool children are frequent not only in

Spanish but also in other Romance languages such as Italian (e.g., Pontecorvo & Zuchermaglio,

1988) and Portuguese (e.g., Martins & Silva, 2001). The syllabic stage is a crucial intermediary

stage in Ferreiro’s theory of spelling development, where it is taken to be the child’s first attempt

to encode the sounds of language. In Romance-speaking countries, Ferreiro’s theory is by far the

predominant paradigm for explaining young children’s spelling acquisition, to such an extent

that early literacy instruction is generally approached as an effort to guide children out of the pre-

syllabic stage of spelling and into the syllabic stage (Silva & Alves-Martins, 2002). In contrast,

the only report we could find of syllabic spellings among English-speaking children was that of

Mills (1998), who studied four- to five-year-old children. A syllabic stage does not play a role in

the prominent theory of Frith (1985), who mentions only a transition from a logographic to an

alphabetic (phonemic) stage. Other researchers have reported that children learning English

rarely or never spell words in a syllabic manner (Kamii et al., 1990; Sulzby, 1985).

It would be surprising if spelling development in Romance languages differs so radically

from spelling development in English. English-speaking children have syllabic awareness and

the ability to manipulate syllables before they can manipulate individual phonemes (Liberman,

Shankweiler, Fischer, & Carter, 1974; Treiman & Zukowski, 1996); and in both languages the

target spelling system is phonemic, with no trace of syllabic orientation. If the syllabic stage of


development is a universal of child development, we would not expect children from different

cultures to develop so differently in the face of substantially similar environmental inputs.

How then could we account for any difference in early spellings between English- and

Portuguese-speaking children? Several types of explanations may have some bearing on the

matter. One tack would be to look for categorical differences in the structure of the languages

themselves. For example, the rhythm of the sentence is said to be based on syllables in the

Romance languages, while it is based on stress in English (Pike, 1945); syllable-based rhythm

may make syllables and vowels more salient for Romance speakers (Kamii et al., 1990). The

syllable structure of English differs from that of Romance languages in allowing multiple

consonants to follow the vowel of the syllable (Blevins, 1995); this may distract attention from

the vowels. Another tack is to look for more graded differences in the languages. Ferreiro and

colleagues (e.g., Ferreiro, 1990) emphasized that English has more one-syllable words than

Spanish, and Kamii et al. pointed to the relative unclearness of many of the unstressed vowels in

English. Such differences may account for somewhat greater attentiveness to vowels in the

Romance languages, but they would not lead to huge, almost categorical, differences in

children’s spellings.

Another approach to explaining cross-language differences in spelling is to look for

differences in the consistency of the mapping from sounds to letters. Vowels in English are much

more inconsistent than consonants (Kessler & Treiman, 2001). Perhaps young English spellers

omit vowels because they are just too difficult. But that would not explain why children omitted

consonants in Romance languages, for consonants are not particularly inconsistent in these

languages. In Portuguese, as in English, vowels tend to be much more difficult to spell than

consonants. For example, /i/ may be spelled i or e, and /u/ may be spelled u or o. Sometimes

contextual cues can help determine how to spell vowels. For example, in Brazilian Portuguese,


/i/ and /u/ are almost always spelled e and o in unstressed final syllables but usually i and u

otherwise. However, beginning spellers may not be aware of these contextual rules, making the

vowel system in Brazilian Portuguese all the more inconsistent. (For an explanation of the

phonetic symbols used in this paper, see International Phonetic Association, 1999.)

If differences in consistency are not sufficient to explain the apparent differences between

the languages, perhaps we should look at differences in the letter name systems (Table 1). In

many countries, children typically learn the names of the alphabet letters at an early age. A

growing body of research shows that young children may adopt a letter name strategy in their

early attempts to spell. The most basic letter name strategy is to take letters as spelling their

entire name. For example, Treiman (1994) found that a nonword such as /vɑɹ/ was frequently

spelled as VR, where R stands for the name of the letter, /ɑɹ/; vowel omission was less common

when the vowel did not form part of a letter name with an adjacent sound. Other data suggest

that beginning spellers are more likely to include vowels in their spellings when the vowel is a

letter name than when it is not (Treiman, 1993). Such findings apply to other languages where

children learn letter names at an early age. Levin, Patel, Margalit, and Barad (2002) found that

Hebrew-speaking Israeli kindergartners were more likely to spell words with letters whose

names were found in the word. Martins and Silva (2001) reported that Portuguese-speaking

children were more likely to use a phonetically accurate letter when the letter’s full name occurs

in the word. In addition, Cardoso-Martins and Batista (2003) have shown that Brazilian

Portuguese-speaking preschoolers often use a phonetically plausible letter in their spelling when

the sound corresponds to a letter name at the beginning of the word.

Given the role of letter names in literacy acquisition, asymmetries in letter name systems

could lead to different patterns in the spellings of young children. Because of differences in letter

names between languages, and because of differences in vocabularies, languages can vary to a


great extent in how many letter names are found in words and in the relative frequency with

which the different letters are so represented. Cardoso-Martins, Resende, and Rodriguez (2002)

began this line of inquiry with Portuguese by showing that, of the 56 most common Portuguese

content words in books for Brazilian kindergartners, 51 contained at least one vowel letter name

in their pronunciation. On a similar list of 56 high-frequency English words, only 17 contained a

letter name. These facts suggest important differences between the languages that may account

for some of the general observations about children’s spelling. For example, perhaps the many

vowel letter names in Portuguese words encourage children to write vowel letters when spelling.

Detailed examination of larger word lists is needed to confirm the generalizability of these

findings and to highlight differences between vowels and consonants. Such large-scale analysis

has so far been limited to English. Working with a list of 6,232 words from kindergarten and

first-grade reading materials (Zeno, Ivenz, Millard, & Duvvuri, 1995), Treiman and Kessler

(2003) found that about 50% of the words in this more general child vocabulary contained a

letter name. In the present research, we extend this line of work to Portuguese.

Statistical properties of the printed words that children see early in life may also influence

the letters that they use. Some evidence for this comes from studies that have investigated

children’s spellings of vowels versus consonants. As discussed previously, Kamii et al. (1990)

suggested that English-speaking kindergartners tend to acquire consonants earlier than vowels.

Similarly, learners of Hebrew tend to represent consonants earlier than vowels (Levin, Share, &

Shatil, 1996). Tolchinsky and Teberosky (1998) found that when children begin to symbolize

phonemes with phonologically appropriate letters, Hebrew speakers showed a preference for

consonants and Spanish speakers for vowels. This difference may be due in part to the types of

letters children most often encounter: In English and especially in Hebrew texts, the proportion

of consonant to vowel letters is very high. Romance languages may have more vowel letters in


text than does English. Mere exposure to different proportions of vowels could be one of the

sources of differences in spelling between English and those other languages.

To quantify and understand some of the differences in children’s spelling between English

and Portuguese, we began in Study 1 by investigating differences in the letter name systems of

English and Portuguese and differences in the letter types—vowels and consonants—

encountered in the languages. Study 2 presents behavioral data on how children’s spelling is

influenced by the differential patterns.

Study 1

Methods

Corpora

We used lists of words drawn from corpora of children’s reading material used for

pedagogical purposes in the respective countries: Pinheiro (1996) for the Brazilian population

and Zeno et al. (1995) for the American population. Because of the smaller sample size of

Pinheiro’s corpus, we selected only words that appear in both the preschool and the first-grade

subcorpora at least once. This criterion allowed us to avoid words that occurred adventitiously in

only a single preschool text. There were 3,621 words in the Brazilian Portuguese list and 6,232

words in the American English list. The larger size of the English list is due in part to the fact

that Zeno et al. do not provide separate counts for preschool and first grade. In any event, we

expect that the spelling and the phonological properties of the word lists in both languages are

representative of preschool vocabulary as a whole. The pronunciations of the Portuguese words

were added by the first author, using pronunciations current in the city of Belo Horizonte. This

accent was chosen because the corpus used in this analysis was drawn from children’s books

used in the schools in this city. Also, the participants in Study 2 were children in Belo Horizonte.

The pronunciations of the English words were taken from the CMU Pronouncing Dictionary


(1998).

Analysis and Results

Letter name analysis. We first calculated the percentage of words that contain at least one

letter name in their pronunciation. We did this for both languages, first over the entire set of

letter names and then for each letter name separately. The analysis was broken down by the

position of the letter name in the word because of the differential salience of those positions. We

do not estimate significance because we computed descriptive statistics over all the relevant

words found in broad and representative corpora, and we are not making predictions about other

words in the language. All the analysis reported here used word type, that is, we did not weight

the words by frequency. However, very similar results are found if the words are weighted by

frequency.

Table 2 shows the results of the letter name analysis. The row labeled “Any” shows that

the percentage of words that have at least one letter name in their pronunciation is twice as great

in Portuguese as in English. Letter names are more common at the ends of words than at the

beginnings. This is particularly true of Portuguese: At the end of words it has almost six times

the percentage of letter names as English. In Portuguese the letter name most commonly found in

words is that of A, followed by I and U. In English, the letter name most commonly found in

words is that of E, followed by A, O, and I. Thus, in both languages, vowel letter names can be

heard in words much more often than consonant letter names.

We also calculated the average number of letter names per word. We performed this

analysis with all the words in the corpora, and we also ran an additional analysis restricted to

two-syllable words stressed on the first syllable. There were 1,243 such words in the Brazilian

Portuguese list and 2,271 in the American English list. Two-syllable words were chosen for this

analysis because we wanted to compare the languages on the basis of words that have a similar


structure in both languages; they are the simplest words that are very common in both English

and Portuguese (Portuguese has relatively few one-syllable words). This procedure also has the

advantage of controlling for word length. We used words with the stress on the first syllable

because that is the most common stress pattern in two-syllable words in both languages.

Table 3 shows the counts broken down by phoneme type, position in the word, and

language. The results indicate that Portuguese has more letter names per word than English. This

difference is due to the number of letter names that are vowels. Portuguese and English words

have similar numbers of consonant letter names, but Portuguese words have 4.3 times as many

vowel letter names as English. At the end of the word, vowel letter names are much more

common than consonant letter names, and in fact Portuguese has extremely few words with

consonant letter names in this position.

To see how helpful letter names may be for children’s initial spelling, we also calculated

their reliability, that is, the percentage of cases in which the letter name heard in the

pronunciation of the word is spelled at least in part with the corresponding letter in the standard

orthography. For example, in English, ceiling contains the letter name /si/, which sound sequence

is partially spelled in that word with the letter that has that name, c. In contrast, the English word

seat has the letter name /si/, but in this word this phoneme string is not spelled with the

corresponding letter, c, but with s. In the former example the consonant letter name is counted as

reliable and in the second it is not. By our definition, a letter name is considered reliable even if

the corresponding letter is only part of a multi-letter spelling. For example, the letter name /e/ is

counted as reliable in make and rain even though additional letters besides a are needed to

accurately spell the vowel sound in those words. We chose a loose criterion because beginning

spellers rarely achieve perfection in such complex spellings, and spelling even the first vowel

letter of a digraph correctly goes a long way toward making children’s spellings understandable


to readers. Another consideration was that Portuguese makes much less use of digraphs than

does English, and no use of final silent e to indicate long vowels; the looser scoring criterion

makes the two languages more comparable.

Table 3 shows the aggregated reliability. The average reliability in English and

Portuguese is similar. Overall, the consonant letter names are more reliable than vowel letter

names in both languages. Vowel letter names at the end of the word are especially unreliable. A

child using a letter name strategy for spelling would be correct only half of the time in

Portuguese and only a little over one third of the time in English. Consonants are reliable most of

the time in both languages. The reliability of consonants at the end of the word is 100%; that is,

every time a consonant letter name is heard at the end of a word, it is spelled with the

corresponding letter. However, consonant letter names are not very common in this position. The

patterns for two-syllable words with the stress on the first syllable are similar to the ones

observed for all the words.

Letter type analysis. We also performed a set of analyses to determine the ratio of vowels

to consonants in the two languages. First, we calculated this ratio in terms of letters in the printed

words, counting y as a vowel. The ratio of vowels to consonants was much higher in Portuguese,

.88, than in English, .58. Next we calculated this ratio in terms of vowel and consonant

phonemes in the corresponding spoken words. These analyses ignored semivowels in both

languages, because of their intermediate status; thus all diphthongs were counted as one vowel.

The results were similar to those for consonant and vowel letters: The vowel–consonant ratio

was .93 for Portuguese and .56 for English. Thus the vowel–consonant ratio is 52% to 66%

higher in Portuguese than in English, depending on whether one counts by letters or by sounds.

Discussion

Study 1 shows that the interaction of the spelling, vocabulary, and letter name systems of


English and Portuguese provides raw material that could lead to differential performance in

spelling tasks. Here we focus on the data that could lead to differences in the generation of

syllabic spellings, in differential preferences for vowels and consonants, and in differential

phonetic plausibility of spellings. The applicability of these data will depend, of course, on the

actual strategies children from the two cultures use when they spell, so detailed discussion of the

implications of these findings will be delayed until after the behavioral data from Study 2 are

presented.

Vowel letters, vowel sounds, and vowel letter names are much more common in

Brazilian Portuguese words than in American English words, confirming the preliminary

observations of Cardoso-Martins et al. (2002). These facts may in part account for the

proliferation of vowels reported in the literature for languages such as Portuguese and Spanish

(e.g., Ferreiro & Teberosky, 1982; Nunes Carraher & Rego, 1984). To the extent that children

rely on a letter name spelling strategy, writing a letter whenever they hear its name in a word, we

would expect them to produce almost two vowels per word, and typically no consonants, in

Portuguese. The same strategy would, for most words, fail entirely in English. To the extent that

children try to reproduce the properties of text they have seen in their environment, one would

expect Brazilian children to write many more vowel letters than U. S. children, simply because

they have seen many more vowels per unit of text. And to the extent that children spell

phonemically using reasonable sound-to-letter correspondences, we would expect Brazilian

children to use more vowels than American children because there are more vowels in

Portuguese than in English words. Thus, the basic statistics of the languages and their writing

systems may explain why beginning spellers of Portuguese seem to use more vowel letters than

beginning spellers of English.

Our analysis may also explain the appearance of syllabic spellings, that is, spellings in


which one letter is written for each syllable. Ferreiro and colleagues (Ferreiro & Teberosky,

1982; Tolchinsky, 2003; Vernon & Ferreiro, 1999) have emphasized the importance of using a

procedure in which children are interviewed once their spelling is completed in order to infer

children’s stage of spelling development. However, if a syllabic stage exists, it should also be

observable in a more objective way in children’s spelling, without the intervention of an

experimenter. In Portuguese, we showed that two-syllable words have, on average, 1.7 letter

names. Thus, if Portuguese-speaking children simply wrote down a letter whose name they hear

its name in a word, they would produce almost one letter (0.84) per syllable. We believe that

children who write 0.84 letters per syllable would generally be accepted as being syllabic. At the

same time, a pure letter name strategy would produce very different results for young speakers of

English: They would produce only 0.33 letters per syllable by this strategy. Thus, a letter name

strategy would lead Brazilian spellers to approximate one letter per syllable but would not lead

U. S. spellers to any such approximation. The letter name hypothesis can account for putatively

syllabic spellings in Portuguese, as well as their general absence in English (Kamii et al., 1990;

Martins & Silva, 2001; Sulzby, 1985), without any reference at all to syllables or to syllable

structure.

Our statistical analyses further implies that the letter name system of Portuguese could be

more helpful to Brazilian children in their initial attempts to spell, compared to that of English.

The fact that the name of a letter can be heard in most Portuguese words may make it easier for

children to gain the insight that writing is a representation of sounds. Once children have gained

that insight, the fact that it can be applied on average two times per word may provide additional

encouragement, as will the fact that teachers and caregivers are more likely to appreciate and

praise spelling attempts where a couple of letters are reasonable representations of sounds in the

word. On these points English, at least with its traditional letter name system, fails badly in


comparison. These factors may help to explain early differences between speakers of English and

other languages in terms of accuracy (e.g., Seymour et al., 2003 for reading). The low

consistency between letters and sounds in English undoubtedly contributes to the relatively low

performance of English speakers in literacy tasks, but the different properties of letter name

systems may also contribute to higher performance in literacy in certain languages than in others.

This may be especially true early in literacy acquisition.

Study 2

Study 2 was designed to address experimentally how spelling of highly similar words

differs between children in the U.S. and Brazil. We asked children to spell words matched on

syllable length—two syllables— but varying in the number of vowel letter names they contain—

one or two. Our first question was whether children in the two cultures indeed differ in their use

of vowel letters. Based on the reports found in the literature, we would expect Portuguese-

speaking children to produce more vowel spellings than English-speaking children, even on

words of the same structure. A second question was whether Portuguese-speaking children show

more evidence of syllabic spellings than U.S. children. Because all our stimuli have two

syllables, we expected Portuguese-speaking children to produce more two-letter spellings than

English speaking children. A third question this study was designed to answer was whether

speakers of the two languages are influenced by vowel letter names in their spelling. We

expected children to use more vowels in the two letter name condition than in the one letter name

condition. We also expected more phonologically plausible spellings in the two letter name

condition than in the one letter name condition.

Method

Participants

Two groups of middle-class children participated. The first group consisted of 49


native speakers of Portuguese (24 males, 25 females) from Belo Horizonte, Brazil. Their ages

ranged from 4 years, 3 months to 6 years, 5 months, with a mean of 5 years, 6 months. The

second group was 43 native speakers of English (22 males, 21 females) from St. Louis in the

United States. Their ages ranged from 4 years, 8 months to 5 years, 10 months, with a mean of 5

years, 3 months. The age range in the Brazilian sample is wider than in the U.S. because of

variations in when children enter preschool in the two countries. Both the U.S. and Brazilian

children were recruited from private preschools in the middle of the school year. Preschools in

both countries did not provide direct instruction in spelling but exposed children to the names

and sounds of the letters of the alphabet, mostly in games and storytelling.

Stimuli

For the spelling task, in the two languages, we used two-syllable words that have

CVCV structure. All words had stress on the first syllable. Only picturable words were chosen so

that they could be presented along with color drawings in order to minimize memory load and

make the task more enjoyable for the children. All the words chosen were expected to be familiar

to children but not to be among the ones that they would be able to spell at the beginning level.

The words were divided into two categories based on the number of letter names in the

pronunciation of the word. The one letter name words, such as bunny /ˈbʌni/, ended in a vowel

letter name (E /i/); the two letter name words, such as pony /ˈponi/, had a vowel letter name in the

first syllable as well (O /o/ and E /i/). We did not use words with no letter names because they are

rare in Portuguese. None of the stimuli contained any consonant letter names (e.g., /bi/ in

English) or consonants for which the typical spelling requires a digraph (e.g., /ʃ/, normally

spelled sh in English). The stimuli were chosen so that the one letter name and two letter name

stimuli had the same number of words with the same final vowel; e.g., each group had the same

number of words ending in /o/. The two groups were also balanced with respect to letter name


patterns for the initial consonant; e.g., each group had the same number of words whose first

consonant’s letter name began with /ɛ/. No word had the same vowel sounds in the first and

second syllables. All stimuli are listed in Table 4.

Procedure

All children were tested individually in a quiet room by an experimenter who was a

native speaker of their language. Each child participated in three sessions on different days. For

most children, the sessions were approximately one week apart. Five children participated in the

two spelling sessions on the same day, with a break between the two tasks. The first session

consisted of pretests evaluating the child’s literacy level: a letter name task, a letter sound task,

and a reading task, in that order. The spelling task was administered in the following two

sessions.

Letter name task. Children were asked to identify letters of the alphabet. The same upper-

case magnetic letters were used in this task that would later be used in the spelling task. This set

of letters comprised the letters necessary for spelling all of the sounds in the words given in the

spelling task: 19 letters for the American children (A, B, D, E, F, G, H, I, K, L, M, N, O, P, R, S,

T, U, and Y) and 17 letters for the Brazilian children (A, B, C, D, E, F, G, I, L, M, O, P, R, S, T,

U, and Z). The letters were placed in front of the child, and the child was asked to choose the

letter that corresponded to the name spoken by the experimenter. The letters were arranged and

queried in a different random order for each child.

Letter sound task. The procedure was very similar to the letter name task, and the same

letters were used. The set of letters was placed in front of the child, and the child was asked to

choose the letter that spells the sound produced by the experimenter. The letter sounds were

presented in random order. We tested the American children on the following sounds: /æ/, /ɑ/,

/b/, /d/, /ɛ/, /f/, /ɡ/, /h/, /ɪ/, /k/, /l/, /m/, /n/, /p/, /ɹ/, /s/, /t/, /ʊ/, and /ʌ/. The Brazilian participants


were tested on the following sounds: /b/, /d/, /e/, /f/, /ɡ/, /h/, /l/, /m/, /o/, /p/, /s/, /t/, and /z/. Those

are the most common sounds of each of the letters given in the spelling task. We omitted testing

for letter sounds that are identical to the letter names already covered in the preceding task.

Reading task. Children were given a word-and-picture reading task. We administered to

the English-speaking sample the same task used by Treiman and Rodriguez (1999), and we

adapted this task for the Portuguese-speaking participants, using words of comparable difficulty

and frequency (alto ‘high’, amarelo ‘yellow’, azul ‘blue’, bola ‘ball’, chuva ‘rain’, comeu ‘ate’,

em ‘in’, eu ‘I’, gato ‘cat’, joga ‘plays’, livro ‘book’, não ‘no’, nós ‘we’, olhe ‘look’, pula

‘jumps’, sou ‘am’, três ‘three’, um ‘one’, vai ‘goes’, vamos ‘let’s go’, verde ‘green’, você ‘you’).

The experimenter showed children 11 different cards with two words and a picture, one card at a

time, and asked the child to identify any items that they knew. If the child did not identify the

items, the experimenter pointed to each item and asked if the child knew it. The order of

presentation of the cards was randomized for each child. The experimenter praised every

response from the child. Children were only scored for their reading of the words; the pictures

were included to make the task less frustrating for children who did not know how to read. All

the words were printed in uppercase letters and were frequent in kindergarten books of the

respective country.

Spelling task. Magnetic letters (the same letters that were presented in the letter-name

task) were placed in random order in front of the children on a magnetic board. The experimenter

said a word while showing a picture representing the same object and then used the word in a

sentence in a way that made reference to the picture. Then the experimenter said the word again

and asked the child to repeat it and spell it. Magnetic letters were used to make the task easier for

young children who do not yet know how to form letters with complete legibility, and to avoid

biasing results in favor of easy-to-draw letters. We told children that we knew that they did not


know how to write yet, but we encouraged them to try their best. The experimenter praised every

response. If children asked for a letter that did not belong to the initial set of magnetic letters or a

letter that repeated a letter that was used already, the experimenter told them to try to spell the

word with the letters provided. If children said they could not spell it without the letter, the

experimenter provided the letter. This happened only rarely. The task was divided into two

sessions of ten words each to minimize effects of fatigue. The order of presentation of the words

was randomized for each child.

Results

Table 5 shows the proportion of correct responses in the letter name, letter sound, and

reading tasks. The results of these pretests are presented as proportions because different

numbers of letters were tested in the two languages. Although most of the children could not

read any words, they knew almost all of the letter names and some letter sounds. The Brazilian

and U.S. groups of children did not differ significantly in letter name knowledge (t(90) = 0.57, p

= .573), letter sound knowledge (t(90) = 0.36, p = .712), or reading knowledge (t(90) = 0.61, p =

.544).

The first analysis examined children’s spellings in terms of the number of letters used,

broken down by whether the letters were consonants or vowels. We counted y as a vowel for

English because none of the words that were presented to children contained the consonantal

sound of y, /j/, while several contained its main vocalic values, /i/ and /aɪ/. Note that 32% of the

letters given to English-speaking children were vowels and 29% of the letters given to

Portuguese-speaking children were vowels. If this slight difference had any effect, we would

expect it to lead to higher vowel usage on the part of the English-speaking children.

Table 6 shows the results broken down by letter name count and language. We used

within-subjects ANOVA with type of letter (consonant or vowel) and letter name count (one or


two) as within-subjects factors, and language (Portuguese or English) as a between-subjects

factor in the by-subject analysis. In the by-item analysis, type of letter was a within-item factor

and condition and language were between-item factors. The results showed a main effect of type

of letter (by subject, F(1,90) = 37.64, p < .001; by item, F(1,36) = 419.80, p < .001): Children

used more consonants than vowels. There was no main effect of letter name count; children did

not differ in the total number of letters used in the two conditions. A main effect of language was

found in the analysis by items (F(1,36) = 19.56, p < .001), but not in the analysis by subjects

(F(1,90) = 0.773, p = .382); Portuguese-speaking children tended to use more letters to spell the

words, but this is probably not a reliable difference. There was an interaction between type of

letter and letter name count (by subject, F(1,90) = 29.05, p < .001; by item, F(1,36) = 34.39, p <

.001). This interaction reflects the fact that children used more vowels to spell words that

contained letter names for each of the vowels, but no difference was found in the use of

consonants in the two conditions. An interaction between type of letter and language was

significant in the by-item analysis (F(1,36) = 34.84, p < .001) and marginally significant by

subjects (F(1,90) = 3.12, p = .081). This interaction shows that Brazilian children used

significantly more vowels than U. S. children, but that the two groups of children did not differ

significantly in their use of consonants. No other effects were found.

To verify whether children had a tendency to misspell vowel letter name sounds with the

corresponding vowel letter, we performed a post hoc analysis of errors made in spelling the last

sounds in the stimuli, which were letter name sounds in all cases. We excluded the 10 stimuli in

which the correct spelling would be the vowel letter in question. In Portuguese, the remaining

words ended with either a /i/ or a /u/, and in English all the words ended with /i/. We examined

all spellings that ended in an incorrect vowel letter. The use of a letter name was the single most

common mistake in both language groups. Children used the named vowel letter in 69% of the


misspellings in English and 70% of the misspellings in Portuguese.

Next, we analyzed whether the first letter of each spelling was a consonant or a vowel.

We focused on the first letter because the beginnings of words are typically highly salient,

especially in our stimuli which had stress on the initial syllable. An initial vowel spelling implies

that children found the first vowel important, omitting the consonant that began all our stimuli.

Portuguese speakers used a vowel letter in the first position of the word in 17.5% of their

spellings, while English speakers did so in only 5.9% of their spellings. To analyze these data,

we used within-subjects ANOVA with letter name count as a within-subject factor and language

as a between-subject factor. In the item analysis we used ANOVA with letter name count and

language as between-item factors and percentage of vowels used in the first position as the

dependent variable. Only the main effect of language was significant (by subjects, F(1,90) =

8.622, p = .004; by items, F(1,36) = 68.38, p < .001).

The use of two letters to represent the words was not very frequent. Two-letter spellings

were produced 10% of the time by the Brazilian children and 19% of the time by the U.S.

children. Only 3 children in each language group produced 80% or more two-letter spellings.

Finally, we categorized the letters in children’s spellings as plausible or not. Plausible

letters are letters that represent the phoneme in ordinary words in the language in question. Table

4 shows which letters we considered plausible for each sound. The vowels /i/ and /u/ in

Portuguese can be spelled either e or i and o or u respectively, and so either letter was counted as

correct for the second vowel of the word. However, for the purposes of this paper, we consider

only i and u as plausible spellings for /i/ and /u/ in the middle position of the words because

otherwise our plausibility count would create an advantage in support of our hypothesis (that

children will show more plausible spellings of vowels in the two letter name condition and in

Portuguese). For each trial, we counted the number of vowel sounds (out of a possible 2) in the


stimulus that had a plausible corresponding letter in the child's spelling; we did the same for the

two consonants in each word. Table 7 reports these as proportions, broken down by language and

the number of letter names in the words. We used within-subjects ANOVA with letter name

count and type of letter (consonant or vowel) as a within-subject factor, and language as a

between-subject factor in the by-subject analysis of these data. In the by-item analysis, type of

letter was a within-item factor and language and letter name count were between-item factors.

We found a main effect of letter name count (by subjects, F(1,90) = 41.20, p < .001; by items,

F(1,36) = 24.70, p < .001) such that children used more plausible spellings for words that had

two letter names. There was also a main effect of language (by subjects, F(1,90) = 4.52, p =

.036; by item, F(1,36) = 66.00, p < .001), meaning that Portuguese speakers made more plausible

spellings than English speakers. We found an interaction between type of letter and letter name

count (by subjects, F(1,90) = 11.10, p < .001; by item, F(1,36) = 9.72, p = .032), such that

children did better on vowels in words with two letter names. This difference was not found with

the consonants. An interaction between type of letter and language was also found (by subjects,

F(1,90) = 38.78, p < .001; by item, F(1,36) = 154.20, p < .001), reflecting the fact that English-

speaking children did better with consonants than vowels and Portuguese-speaking children

showed the reverse pattern.

Discussion

Study 2 was designed to investigate whether English and Portuguese-speaking children

differ in their spellings of words that are substantially equated across languages. A first question

was whether children from the two cultures really differ in whether they favor vowel or

consonant letters when the number of consonant and vowel sounds is controlled. We found that

the Brazilian children did use more vowel letters than did the U.S. children. It was particularly

striking that the Brazilian children were three times more likely to start a word with a vowel


letter than the U.S. children. This difference cannot be explained by the stimuli used in the

experiment because they all had the same number of vowel and consonant sounds. However, the

difference may be explained by the differing properties of the writing systems to which children

have been exposed. As found in Study 1, children learning to spell in Portuguese have been

exposed to printed words that have an almost one-to-one vowel–consonant ratio. That could

explain the fact that Portuguese-speaking children used vowels much more frequently than

English-speaking children. Children who see written Portuguese may pick up on the fact that

words have a high proportion of vowel letters in them. Conceivably, the preponderance of

vowels disproportionately attunes children to attend to vowels in spelling. Taken together, this

evidence provides an explanation for reports in the literature of all-vowel spellings by

Portuguese-speaking children and of consonant spellings by English-speaking children (e.g.,

(Nunes Carraher & Rego, 1984; Kamii et al., 1990).

A second question was whether children from the two cultures differ in their use of

syllabic spellings. The literature (e.g., Ferreiro & Teberosky, 1982) makes very frequent

reference to syllabic spellings among Romance-speaking children in the age group under

consideration, but evidence for this in our data is scant. Because all of our stimuli had two

syllables, syllabic spellers would be expected to produce two-letter spellings in all responses. But

only a small fraction of the spellings were two letters in length (10% of the Portuguese

spellings), and, contrary to all expectation, such spellings were almost twice as common in

English. Altogether, there is only weak evidence for a syllabic stage of development, and

certainly no evidence that Portuguese speakers evince this tendency more than English speakers.

A third issue addressed by Study 2 was whether children are influenced by letter names

in their spelling. All our analyses showed that the answer to this question is affirmative. In both

English and Portuguese, the presence of vowel letter names in the words led children to use more


vowels in their spellings. Our post hoc analysis of vowel spelling errors at the end of words also

showed heavy use of letters for their letter-name value, as when children spelled pony with a

final E; use of a letter name strategy is very plausibly a major contributor to that effect. Letter

names also made spellings more phonologically accurate. This accords with previous findings

(Cardoso-Martins & Batista, 2003; Martins & Silva, 2001) that Portuguese-speaking children are

more likely to spell phonemes with a phonetically plausible letter when that phoneme constitutes

the letter’s name. Our results show that vowel letter names do influence children’s spelling,

eliciting both more vowel spellings and more accurate phonological spellings of the stimuli.

General Discussion

Previous studies have suggested some potentially important differences between speakers

of Romance languages and speakers of English in how they approach the task of learning to

spell. In particular, speakers of Romance languages appear to use more vowels and have been

reported to spell syllabically, writing one symbol per syllable. Given the impact that such

differences may have for theories of spelling acquisition—for example, whether children pass

through a syllabic stage of development—and for the teaching of literacy, it is important to study

them quantitatively.

We found some statistically significant differences in the spelling productions of

Brazilian and U.S. children, but these are not of the magnitude that would suggest categorical

differences in how children approach spelling in the different countries. Instead of great

categorical differences between children’s spellings in the two languages, we found small but

significant quantitative differences. We believe they are the sorts of differences best accounted

for by graded differences in environmental input rather than by large qualitative differences

between the languages. One quantitative difference in the input is that Portuguese words have

more vowels than English words, as documented in Study 1. Given this, it should not be


surprising to see more vowels in the spontaneous spellings of Portuguese speakers, as we did in

Study 2. Even when the words that English and Portuguese children are asked to spell contain an

identical number of vowels, as in that study, the relative preponderance of vowels in Portuguese

vocabulary as a whole seems to exert a small but measurable indirect influence. When young

spellers are uncertain as to which letter to use, their last resort may well be to fall back on

familiarity. Assuming, as is almost certain for middle-class children, that they have seen

substantial amounts of text in their environment, young children may favor letters they have seen

more often. Because vowel letters are disproportionately more common in Portuguese text than

in English text, Brazilians would be more likely to select a vowel letter than U.S. children would

be.

A letter name spelling strategy may also contribute to the greater number of vowel letters

in early Portuguese spelling than in English spelling. Study 2 documented that children are more

likely to insert a vowel letter in their spelling if they hear its letter name in a word. While the

magnitude of that letter name effect is not significantly greater per se in Portuguese than in

English, one must consider an interaction with the vocabulary of the respective languages.

Children’s vocabulary in Portuguese contains more than four times the number of vowel letter

names per word. Therefore, a letter name effect will lead to proportionately more vowel spellings

in Portuguese than in English when spelling words that are typical of the respective languages.

The differences we have documented between Portuguese and English in frequency of

vowel letters and letter names could lead to spellings that appear to support the idea of a syllabic

stage for Portuguese. Spellings such as UUU for urubu are often reported as demonstrations of

syllabic spellings (Nunes Carraher & Rego, 1984), in that the child represented each syllable

with one letter. However, we have shown that children are more likely to spell letters when their

name appears in a word. Children who rely on that strategy to any great extent should show a


strong tendency to spell vowels and omit consonants in Portuguese; the effect would presumably

be stronger in Spanish and in most accents of Italian, where every vowel sound is also a letter

name. The end result of such a letter name strategy—spelling letters whose names are heard in

words, i.e., mostly just the vowel of each syllable—is indistinguishable or easily confounded

with a syllabic spelling strategy—writing a single symbol to characterize each syllable. On the

syllabic spelling hypothesis, though, it is difficult to explain why examples of syllabic spellings

almost invariably consist of strings of vowel letters to the exclusion of consonant letters. Why

should consonant letters be less capable than vowel letters of representing a syllable in the

child’s mind? Our alternative hypothesis—that the apparent syllabic spellings in Portuguese and

other languages actually reflect use of letter names—explains this phenomenon. Spellings that

contain a large number of vowels can easily be accounted for by letter name spelling strategy,

which is already required to account for other documented phenomena. We thus question the

need for the complicating assumption of a syllabic spelling stage, an assumption which is

generally held among researchers studying Romance languages and which is widely applied by

preschool educators in many Latin American and European countries (e.g., Martins & Silva,

2001).

Finally, our results may have bearing on why English-speaking children often perform

relatively poorly in literacy tasks. Cross-linguistic studies (Aro & Wimmer, 2003; Defior et al.,

2002; Seymour et al., 2003) have emphasized letter–sound consistency as the major factor.

Although this is undoubtedly true across the spellers’ lifespan, we believe that other factors may

be rather important as children just begin to learn how to write. At early ages, it is more

appropriate to hope for phonetically plausible spellings rather than conventionally perfect ones,

and so consistency in the standard orthography is not really a relevant measure. Our data show

that English-speaking children produce more plausible spellings for consonants than vowels,


while Portuguese-speaking children show the reverse pattern. We propose that, at least for young

children, those differences could be explained not only by consistency but also by other

statistical properties that vary across languages, especially the letter name system and its

interaction with children’s vocabulary. Admittedly, the differences uncovered in our study are of

small magnitude. However, considering that children in the present study were very young,

leading to random noise in the data, and that the words presented to the two language groups

were controlled for number of letter names, we believe that the numbers are important indicators

of language differences. These differences would likely be boosted in a naturalistic setting, when

Portuguese- speaking children will encounter many more words that contain vowel letter names

than English-speaking children. The fact that children can profit from a letter name strategy more

often in Portuguese than in English, especially with respect to vowels, may exert a strong

influence on young spellers, leading to greater overall reliance on and more plausible production

of vowels in Portuguese-speaking children.

Cross-cultural studies are important for determining how specific properties of a

language can make easier or harder the task of children who are trying to read and spell. The

studies conducted in this paper have quantified differences between the writing systems of

English and Portuguese, expanding previous work by analyzing aspects of the language that have

been neglected by past cross-linguistic studies: letter names and letter patterns. Our results

demonstrate how those differences may lead to different patterns in performance of children

acquiring spelling in those writing systems and shed doubt on the commonly accepted view that

learners of Portuguese and other Romance languages go through a syllabic stage in the

development of spelling.


References

Aro, M. & Wimmer H. (2003). Learning to read: English in comparison to six more regular

orthographies. Applied Psycholinguistics, 24, 621–635.

Blevins, J. (1995). The syllable in phonological theory. In J. A. Goldsmith (Ed.), The handbook

of phonological theory (pp. 206–244). Cambridge, MA: Blackwell.

Cardoso-Martins, C. & Batista, A.C.E. (April, 2003). The role of letter name knowledge in

learning to connect print to speech: Evidence from Brazilian Portuguese-speaking

children. Paper presented at the meeting of the Society for Research in Child

Development, Tampa, FL.

Cardoso-Martins, C., Resende, S. M., & Rodrigues, L. A. (2002). Letter name knowledge and

the ability to learn to read by processing letter–phoneme relations in words: Evidence

from Brazilian Portuguese-speaking children. Reading and Writing, 10, 10–20.

CMU Pronouncing Dictionary (version 0.6) [Data file]. (1998). Carnegie Mellon University,

Speech at CMU Web site: http://ftp.cs.cmu/project/speech/dict/cmudict.0.6.

Defior, S., Martos, F., & Cary, L. (2002). Differences in reading acquisition development in

shallow orthographies: Portuguese and Spanish. Applied Psycholinguistics, 23, 135–148.

Ferreiro, E. (1990). Literacy development: Psychogenesis. In Y. Goodman (Ed.), How children

construct literacy: Piagetian perspectives (pp. 12–25). Newark, DE: International

Reading Association.

Ferreiro, E. & Teberosky, A. (1982). Literacy before schooling. New York: Heinemann.

Frith, U. (1985). Beneath the surface of developmental dyslexia. In K. E. Patterson, J. C.

Marshall, & M. Coltheart (Eds.), Surface dyslexia: Neuropsychological and cognitive

studies of phonological reading (pp. 301–330). London: Erlbaum.

International Phonetic Association (1999). Handbook of the International Phonetic Association:


A guide to the use of the International Phonetic Alphabet. Cambridge, England:

Cambridge University Press.

Kamii, C., Long, R., Manning, M., & Manning, G. (1990). Spelling in kindergarten: A

constructivist analysis comparing Spanish-speaking and English-speaking children.

Journal of Research in Childhood Education, 4, 91–97.

Kessler, B., & Treiman, R. (2001). Relationships between sounds and letters in English

monosyllables. Journal of Memory and Language, 44, 592–617.

Levin, I., Patel, S., Margalit, T., & Barad, N. (2002). Letter names: Effect on letter saying,

spelling, and word recognition in Hebrew. Applied Psycholinguistics, 23, 269–300.

Levin, I., Share, D. L., & Shatil, E. (1996). A qualitative–quantitative study of preschool writing:

Its development and contribution to school literacy. In C. M. Levy & S.

Ransdell (Eds.), The science of writing: Theories, methods, individual

differences, and applications (pp. 271–293). Mahwah, NJ: Erlbaum.

Liberman, I.Y., Shankweiler, D. Fischer, F.W., & Carter, B. (1974). Explicit syllable and

phoneme segmentation in the young children. Journal of Experimental Child Psychology,

18, 201–212.

Martins, M. A., & Silva, C. (2001). Letter names, phonological awareness and the phonetization

of writing. European Journal of Psychology of Education, 4, 605–617.

Mills, L. (1998). Syllabic stage of English speaking preschoolers. Reading Research and

Instruction, 37, 297–317.

Nunes Carraher, T. & Rego, L. R. B. (1984). Desenvolvimento cognitivo e alfabetização. Revista

Brasileira de Estudos Pedagógicos, 63, 38–55.

Pinheiro, A. M. V. (1996). Contagem de freqüência de ocorrência de palavras expostas a

crianças na faixa pré-escolar e séries iniciais do 1o. grau. São Paulo, Brazil: Associação


Brasileira de Dislexia.

Pike, K. L. (1945). The intonation of American English. Ann Arbor, MI: University of Michigan

Press.

Pontecorvo, C. & Zucchermaglio, C. (1988). Modes of differentiation in children’s writing

construction. European Journal of Psychology of Education, 3, 371–384.

Read, C. (1986). Children’s creative spelling. London: Routledge & Kegan Paul.

Seymour, P. H. K., Aro, M., & Erskine, J. M. (2003). Foundation literacy acquisition in

European languages. British Journal of Psychology, 94, 143–175.

Silva, C. & Alves-Martins, M. (2002). Phonological skills and writing of pre-syllabic children.

Reading Research Quarterly, 37, 466–483.

Sulzby, E. (1985). Kindergartners as writers and readers. In M. Farr (Ed.), Advances in writing

research, Volume 1. Children’s early writing development (pp 127–199). Norwood, NJ:

Ablex Publishing Corporation.

Tolchinsky, L. (2003). The cradle of culture and what children know about writing and numbers

before being taught. Mahwah, NJ: Lawrence Erlbaum.

Tolchinsky, L., & Teberosky, A. (1998). The development of word segmentation and writing in

two scripts. Cognitive Development, 13, 1–24.

Treiman, R. (1993). Beginning to spell: A study of first-grade children. New York: Oxford

University Press.

Treiman, R. (1994). Use of consonant letter names in beginning spelling. Developmental

Psychology, 30, 567–580.

Treiman, R. & Kessler, B. (2003). The role of letter names in the acquisition of literacy. In R.

Kail (Ed.), Advances in Child Development and Behavior, Vol. 31 (pp. 105–135). San

Diego, CA: Academic Press.


Treiman, R. & Rodriguez, K. (1999). Young children use letter names in learning to read words.

Psychological Science, 10, 334–338.

Treiman, R. & Zukowski, A. (1996). Children’s sensitivity to syllables, onsets, rimes, and

phonemes. Journal of Experimental Child Psychology, 61, 193–215.

Varnhagen, C. K., Boechler, P. M., & Steffler, D. J. (1999). Phonological and orthographic

influences on children’s vowel spelling. Scientific Studies of Reading, 4, 363–379.

Vernon, S. A. & Ferreiro, E. (1999). Writing development: A neglected variable in the

consideration of phonological awareness. Harvard Educational Review, 69, 395–414.

Zeno, S. M., Ivenz, S. H., Millard, R. T., & Duvvuri, R. (1995). The educator’s word frequency

guide. Brewster, NY: Touchstone Applied Science Associates.


Author Note

Tatiana Cury Pollo, Psychology Department, Washington University in St. Louis; Brett

Kessler, Psychology Department; Washington University in St. Louis; Rebecca Treiman,

Psychology Department; Washington University in St. Louis. This work was supported in part

by Grant BCS-0130763 from the National Science Foundation. We thank Cláudia Cardoso-

Martins and Mirelle França Michalick for their help with the Brazilian part of the study and

Annukka Lehtonen, Amanda Mount, Sabria McElroy, and Angela Wong for helping with the

American part of the study. We are also grateful to the faculty, parents, and children from the

schools who participated in the study: Colégio Marista Dom Silvério, Forsyth School, and

Wilson School. Correspondence may be addressed to Tatiana Pollo, Campus Box 1125, One

Brookings Drive, St. Louis MO 63130 USA. Email: [email protected]


Table 1

Letter Names in Brazilian Portuguese and American English

Letter Brazilian Portuguese American English

A /a/ /e/ B /be/ /bi/ C /se/ /si/ D /de/ /di/ E /ɛ/ /i/ F /ˈɛfi/ /ɛf/ G /ʒe/ /dʒi/ H /aˈɡa/ /etʃ/ I /i/ /aɪ/ J /ˈʒɔta/ /dʒe/ K — /ke/ L /ˈɛli/ /ɛl/ M /ˈemi/ /ɛm/ N /ˈeni/ /ɛn/ O /ɔ/ /o/ P /pe/ /pi/ Q /ke/ /kju/ R /ˈɛhi/ /ɑɹ/ S /ˈɛsi/ /ɛs/ T /te/ /ti/ U /u/ /ju/ V /ve/ /vi/ W — /ˈdʌblju/ X /ʃis/ /ɛks/ Y — /waɪ/ Z /ze/ /zi/

Note. Letters that are not part of the core alphabet of Portuguese are marked with a dash.


Table 2

Percentage of Words that Contain at Least One Letter Name in their Pronunciation

Anywhere Beginning End

Letter Portuguese English Portuguese English Portuguese English

Any 95.3 46.9 20.9 6.1 71.6 12.1

A 61.5 9.5 8.3 0.5 34.5 0.8

B 0.9 0.6 0.5 0.4 0.1 0.2

C 2.7 1 1.6 0.5 0.3 0.4

D 2.5 1.1 1.5 0.1 0.1 0.7

E 8 16.5 0.4 0.4 0.6 7.9

F 0 0.1 0 0 0 0

G 0.6 0.2 0.4 0.1 0 0.1

H 0 0 0 0 0 0

I 38.8 7.6 3.6 0.4 10.5 0.5

J 0 0.2 0 0.2 0 0

K — 0.7 — 0.3 — 0

L 0.2 1.8 0 0.2 0 0.3

M 0.1 0.6 0 0.2 0.1 0

N 0.1 2.6 0 0.5 0 0.3

O 7.6 7.6 0.5 0.6 0.4 1.6

P 1.8 0.9 1.1 0.4 0 0.2

Q 0.7 0.1 0.2 0.1 0.1 0

R 0 2.9 0 0.3 0 0.2

S 0.2 1.2 0 0 0.2 0.3

T 1.4 1.2 0.6 0.2 0.1 0.7

U 37.9 1 1.2 0.3 24.4 0.1

V 1.8 0.1 0.9 0 0.2 0.1

W — 0 — 0 — 0

X 0 0.2 0 0.1 0 0

Y — 0.5 — 0.4 — 0

Z 0.8 0.4 0.1 0 0.1 0.2


Table 3

Average Number of Letter Names per Words and their Reliability in English and Portuguese

(Results For the Two Syllable Words in Parentheses)

Portuguese English

Type of letter

Letter names

per word

Reliability

(%)

Letter names

per word

Reliability

(%)

Anywhere in word

Any Letter 2.00 (1.68) 76.0 (70.2) 0.59 (0.66) 79.6 (74.2)

Consonants 0.14 (0.06) 85.7 (85.7) 0.16 (0.18) 87.5 (83.3)

Vowels 1.86 (1.62) 75.3 (69.8) 0.43 (0.48) 76.7 (79.2)

Beginning of word

Any Letter 0.21 (0.12) 81.0 (91.7) 0.06 (0.06) 83.3 (83.3)

Consonants 0.07 (0.06) 85.7 (83.3) 0.04 (0.04) 75.0 (75.0)

Vowels 0.14 (0.06) 78.6 (100.0) 0.02 (0.02) 100.0 (100.0)

End of word

Any Letter 0.72 (0.83) 54.2 (49.2) 0.15 (0.23) 53.3 (43.5)

Consonants 0.02 (0.00) 100.0 (—) 0.04 (0.06) 100.0 (83.3)

Vowels 0.70 (0.66) 52.8 (49.4) 0.11 (0.17) 36.4 (29.4)


Table 4

Stimuli and Letters Considered Plausible Representations of Sounds

One letter name stimuli

Two letter name stimuli

Word C1 V1 C2 V2 Word C1 V1 C2 V2

English

bully /ˈbʊli/ B U/O L E/Y bony /ˈboni/ B O N E/Y

bunny /ˈbʌni/ B U N E/Y goalie /ˈɡoli/ G O L E/Y

dolly /ˈdɑli/ D O/A L E/Y holey /ˈholi/ H O L E/Y

gummy /ˈɡʌmi/ G U M E/Y Nike /ˈnaɪki/ N I/Y K E/Y

hippo /ˈhɪpo/ H I P O phony /ˈfoni/ F O N E/Y

holly /ˈhɑli/ H O/A L E/Y pony /ˈponi/ P O N E/Y

limo /ˈlɪmo/ L I M O rainy /ˈɹeni/ R A N E/Y

Mickey /ˈmɪki/ M I K E/Y rhino /ˈɹaɪno/ R I/Y N O

rocky /ˈɹɑki/ R O/A K E/Y silo /ˈsaɪlo/ S I/Y L O

sunny /ˈsʌni/ S U N E/Y tiny /ˈtaɪni/ T I/Y N E/Y

Portuguese

bolo /ˈbolu/ B O L O/U bota /ˈbɔta/ B O T A

boto /ˈbotu/ B O T O/U bote /ˈbɔti/ B O T E/I

doce /ˈdosi/ D O S/C E/I bule /ˈbuli/ B U L E/I

fogo /ˈfoɡu/ F O G O/U figo /ˈfiɡu/ F I G O/U

gota /ˈɡota/ G O T A fila /ˈfila/ F I L A

mesa /ˈmeza/ M E S/Z A gado /ˈɡadu/ G A D O/U

rede /ˈhedi/ R E D E/I mula /ˈmula/ M U L A

remo /ˈhemu/ R E M O/U ralo /ˈhalu/ R A L O/U

rolo /ˈholu/ R O L O/U rato /ˈhatu/ R A T O/U

sopa /ˈsopa/ S/C O P A sapo /ˈsapu/ S/C A P O/U


Table 5

Proportion of Correct Responses in Pretests (Standard Deviations in Parentheses) of Study 2

Pretest measure Portuguese English

Letter name task .98 (.07) .97 (.06)

Letter sound task .72 (.26) .70 (.20)

Reading task .17 (.28) .20 (.26)


Table 6

Mean Number of Letters Used to Spell Words in Each Condition (Standard Deviations in

Parentheses) in Study 2

Portuguese

English

Letter name count Vowel Consonant Total Vowel Consonant Total

One letter name 1.69 (0.41) 2.32 (0.94) 4.01 (1.01) 1.42 (0.65) 2.30 (1.20) 3.72 (1.70)

Two letter names 1.90(0.33) 2.11 (0.86) 4.02 (1.01) 1.58 (0.57) 2.21 (1.37) 3.79 (1.77)

All stimuli 1.79 (0.32) 2.22 (0.87) 4.02 (0.98) 1.50 (0.59) 2.25 (1.26) 3.75 (1.73)


Table 7

Proportion of Plausible Representations of Consonants and Vowels (Standard Deviations in

Parentheses) in Study 2

Portuguese

English

Type of

phoneme

One letter

name

Two letter

names

All

One letter

name

Two letter

names

All

Consonant .61 (.30) .65 (.30) .63 (.30) .66 (.27) .68 (.26) .67 (.26)

Vowel .71 (.26) .83 (.21) .77 (.22) .49 (.23) .57 (.23) .53 (.22)

Any .66 (.26) .74 (.23) .70 (.24) .58 (.21) .63 (.21) .60 (.21)

Documents

Running Head: CHILDREN’S SPELLING IN ENGLISH AND PORTUGUESEspell.psychology.wustl.edu/PolloVSyllLN/PolloVSyllLN.pdf · Spanish, and Kamii et al. pointed to the relative unclearness